Abstract
• Background and Aims Sex changes within the genus Acer (Aceraceae) may occur because of associations of sex expression and plant health. In this study, a natural population of Acer rufinerve was monitored to clarify the sex change patterns, the relationship between sex expression and plant health, and the causal environmental conditions that precede sex changes.
• Methods Sex expression, growth rate and mortality of A. rufinerve trees in a natural population were monitored from 1992 to 1997.
• Key Results Three types of sex expression were observed among A. rufinerve: male, female and bisexual. Among the three types of sex expression, sex changes occurred in all directions. In the growing season of 1994, precipitation was reduced. Stem growth rate decreased and mortality was high in 1994. In the spring of 1995, a drastic sex change from male to female or to bisexual occurred. As a result, the sex ratio became female‐biased in 1995, although it had been male‐biased from 1992 to 1994. In 1996 and 1997, the proportion of males in the population increased, partly as a result of female mortality and partly as a result of female‐to‐male sex changes. Sex expression of A. rufinerve was associated with their growth rate and mortality. The growth rate decreased for trees whose sex changed from male to female or to bisexual, and increased for trees whose sex changed from female to male or to bisexual. Dead trees reproduced as females before they died, except for those that died as males in 1994.
• Conclusions One explanation for the sex change towards increasing femaleness for this A. rufinerve population in 1995 was the deterioration of plant health in the previous growing season, because of reduced precipitation. Sex changes of unhealthy and dying A. rufinerve towards femaleness may facilitate re‐occupancy by offspring in gaps created by the death of A. rufinerve trees.
Key words: Acer rufinerve, environmental conditions, growth rate, meteorological conditions, mortality, sex change, plant health, sex expression, reproductive cost, sex ratio, tree species
INTRODUCTION
Sex changes in plants have been considered a strategy of sex allocation, to enhance fitness over the lifetime of the plants (Policansky, 1981). The ability of plants to change their sex expression is favoured by natural selection when the reproductive success of an individual as a male or female is influenced by size, age or environmental conditions (Ghiselin, 1969; Charnov and Bull, 1977; Freeman et al., 1980).
A widely accepted explanation for sex change is offered by the size advantage model (Ghiselin, 1969; Charnov, 1982), which predicts that plants change their sex expression when size‐ or age‐specific reproductive success differs between males and females. For several plant species, including some herbs and trees, sex expression and plant size are correlated, i.e. small plants reproduce as males while larger plants reproduce as females (Bierzychudek, 1982; Lovett Doust and Cavers, 1982; Kinoshita, 1986; Schlessman, 1991; Yamashita and Abe, 2002). Sex changes from male to female with increased individual size have also been demonstrated in natural populations (Bierzychudek, 1982; Kinoshita, 1987). Male to female sex changes have been attributed to the internal resource status of plants, which is associated with their size or age (Lloyd and Bawa, 1984; Kinoshita, 1986), because reproductive costs often differ between sexes (Lloyd and Webb, 1977; Korpelainen, 1994; Abe, 2002).
Freeman et al. (1980) suggested that sex changes in several plants occur as a response to varying ambient environments. This theory is supported by the correlation between environmental conditions and sex expression of plants. Males are more common in harsher environments, such as xeric, nutrient‐poor or shaded conditions (Freeman et al., 1976; Lloyd and Bawa, 1984; Zimmerman, 1991; Ortiz et al., 2002). McArthur (1977) monitored a population of Atriplex canescens and reported that the number of females decreased as a result of sex changes following stressful meteorological conditions.
In this study, sex change patterns in Acer refinerve (Aceraceae) were investigated. Several researchers have reported sex changes within the genus Acer (DeJong, 1976; Sakai, 1978, 1990; Hibbs and Fischer, 1979; Barker et al., 1982; Primack and McCall, 1986; Matsui, 1995; Ushimaru and Matsui, 2001) and, according to previous studies, sex expression in Acer species appears to be dependent on the ambient environment rather than on size, although bisexuals were significantly larger than males or females in Acer saccharinum (Sakai, 1978; Sakai and Oden, 1983). Barker et al. (1982) recorded the sex expression of A. grandidentatum trees over two consecutive reproductive seasons and found more males after dry meteorological conditions than after wet conditions. Hibbs and Fischer (1979) reported that sex expression of A. pensylvanicum showed no dependency on plant height or age and that females were less healthy than males. They suggested that an individual A. pensylvanicum starts reproduction as a male and then becomes female after a deterioration of growth conditions that may be induced by progressive crown closure in a forest. Matsui (1995) monitored sex expression in a population of A. rufinerve for 5 years. He reported that the directional sex changes predicted by the size advantage model were not observed and that unhealthy plants, which may have been suffering from water stress, tended to be females. Hibbs and Fischer (1979) and Matsui (1995) proposed a hypothetical mechanism that plant health governs the sex expression of plants, and predicted that environmental conditions would affect plant health. To test this hypothetical mechanism, temporal relativity or causality between environmental conditions, plant health and sex expression need to be clarified.
The objectives of this study were: (a) to clarify the sex change patterns of A. rufinerve; (b) to evaluate the relationship between sex expression and health of A. rufinerve trees; and (c) to detect the causal environmental conditions that precede sex changes. To achieve these, sex expression in a natural population of A. rufinerve was monitored over 6 years and the growth rate and mortality of trees, as indicators of health, were measured. An attempt was made to relate the health of A. rufinerve to meteorological conditions. The advantages of sex changes for A. rufinerve in varying environments are discussed.
MATERIALS AND METHODS
Study species
Acer rufinerve Sieb. et Zucc. is a deciduous tree endemic to Japan (Ogata, 1999). Leaves are opposite, shallowly 3–5‐lobed, and 8–15 cm long and wide. Acer rufinerve is an early successional species that occupies open habitats (Sakai, 1987) such as forest edges and canopy gaps. Flowering starts in April–May synchronously with leaf emergence. An inflorescence consists of a raceme with approx. 15 flowers and is 5–10 cm long. Pollen is dispersed by small bees and dipterans (Matsui, 1991); samaras ripen in autumn and are dispersed by wind. Sex changes in A. rufinerve have been observed by Matsui (1995) and Ushimaru and Matsui (2001). Matsui (1995) reported three types of sex expression in A. rufinerve: male plants with only male inflorescences, female plants with only female inflorescences, and bisexual plants with both male and female inflorescences. Stem diameter growth of A. rufinerve starts early in May, and it stops early in October (S. Nanami, unpubl. res.).
Study site
The study site was a forested area located on Mt Mikasa (294 m altitude, 34°41′N, 135°51′E), Nara City, Japan. The vegetation on Mt Mikasa is protected from human disturbances and is dominated by the evergreen conifer, Podocarpus nagi (Podocarpaceae), and the evergreen broad‐leaved tree Neolitsea aciculata (Lauraceae); deciduous broad‐leaved tree species, including A. rufinerve, are mixed among these dominant species (Suganuma and Kawai, 1978; Ohmae et al., 1996; Nanami et al., 1999; Tateno and Kawaguchi, 2002).
The Nara Meteorological Station (104 m a.s.l.), which is located about 2·5 km north‐west of our study forest, recorded an average annual temperature and precipitation of 14·4 °C and 1354·7 mm between 1961 and 1990. During the 6 years of our study, from 1992 to 1997, a drought occurred in 1994 (Anon., 1994). Precipitation from April to November, the growing season of A. rufinerve, in 1992, 1993, 1995, 1996 and 1997 ranged from 86 to 119 % of the average recorded from 1961 to 1990, whereas in 1994 it was only 49 % (Fig. 1).
Fig. 1. Precipitation from April to November, the growing season of Acer rufinerve, between 1992 and 1997 at the Nara Meteorological Station (104 m a.s.l.), about 2·5 km north‐west of our study forest on Mt Mikasa, Nara City, Japan. * The data of 1997 does not include precipitation in May due to a loss of measurement. ** Average precipitation 1961–1990.
Field methods
In 1988, a 1·48‐ha plot (40 × 370 m) was established from the base to the peak of Mt Mikasa. The plot was located on a north‐west facing slope with an inclination of about 20° and an elevation range of 130 m from the lowest to the highest point. The elevation increased monotonically along the slope with no complex topography. The plot was divided into 5 × 5‐m quadrats. Within each quadrat, the x‐ and y‐coordinates of the centre of each plant ≥5 cm in stem diameter at breast height (dbh, 130 cm above ground level) was mapped, and the species and size were recorded.
In March 1992, 58 A. rufinerve trees were counted within the plot. The sex of each A. rufinerve tree ≥5 cm in dbh was determined by observing inflorescences in April–May and fruiting through binoculars every year from 1992 to 1997. When the trees were observed for flowering they were checked carefully to judge if they had died or not. Mortality of each tree in a given year was defined as the mortality between flowering in the given year and flowering in the next year. If a tree was dead when the flowers were observed, it was determined that the tree had died in the previous year. Also in the spring of 1998, the status of the trees was checked to judge if the tree had died or not. In March 1992, an aluminium band‐type dendrometer (Liming, 1957) was installed at breast height on each tree. This instrument consists of a band of aluminium that encircles the tree trunk and is held in place by a coil spring. The diameter growth of each trunk was measured annually, based on changes in the aluminium bands.
Data analysis
Any deviation of sex ratios (number of males : number of females) from 1 : 1 was tested using χ2 tests. The frequency of sex changes was compared among the six study years by Fisher’s exact test with a Bonferroni correction.
To estimate the health of trees in each year at the population level, the average absolute growth rate of stem diameter (AGR) and the mortality of trees between the 6 years were compared. To estimate the health of trees in each year at the individual level, the AGR of a tree in a given year was compared with the AGR in the previous year using a Wilcoxon test. To test the dependency of sex expression on tree size, dbh was compared among sexes by a Kruskal–Wallis test.
To analyse the association of mortality and sex expression, mortality was compared among males, females and bisexuals. Furthermore, the growth rates of individual trees were compared, before and after a sex change, for individuals that had changed their sex in April of 1994 and 1995, as a sufficient number of sex changes had been observed in these 2 years to support statistical analysis (see Results).
RESULTS
Sex ratio and size structure
The ratio of sex expression among A. rufinerve varied from year to year (Fig. 2). Males outnumbered females from 1992 to 1994, and the male : female ratio was significantly male‐biased in 1993 (male : female ratio = 2·62, χ2 = 9·38, P < 0·01) and 1994 (male : female ratio = 3·50, χ2 = 13·89, P < 0·001). In 1995, the sex ratio became significantly female‐biased (male : female ratio = 0·39, χ2 = 6·13, P < 0·05). From 1996 to 1997, the proportion of males in the population increased, and males outnumbered females again in 1997, although a significant bias was not observed.
Fig. 2. Relative frequency of male, bisexual, female and non‐flowering Acer rufinerve in each year from 1992 to 1997. Sample size is indicated above each bar. The values in parentheses are male : female ratios, and the significance of any deviations from the expected 1 : 1 sex ratio are indicated (χ2 test); * P < 0·05; ** P < 0·01; *** P < 0·001; NS, P ≥ 0·05.
The dbh of trees did not differ among males and females in any year (Fig. 3). However, bisexual plants were significantly larger than either males or females in 1992 and 1993 (Fig. 3).
Fig. 3. Stem diameter at breast height (dbh) of male, bisexual and female Acer rufinerve in each year from 1992 to 1997. The bars represent the median dbh for each sex. Median dbh values differed significantly among males, females and bisexuals (P < 0·05, Kruskal–Wallis test) in 1992 and 1993. Values denoted by the same letters are not significantly different from each other at P < 0·05 (Mann–Whitney test with a Bonferroni correction).
Sex change
Sex changes were observed in all directions among males, females and bisexuals (Table 1). Out of 28 trees which survived in 1997, six trees were consistently male (two trees had a year of non‐flowering) and two trees were consistently female throughout the six observations. The frequency of sex change varied from once to three times among the trees that changed their sex.
Table 1.
Sex change sequences among male (M), female (F), bisexual (B) and non‐flowering (N) from 1992 to 1997 of the 58 Acer rufinerve trees studied
| Year | Year of | No. of | |||||
| 1992 | 1993 | 1994 | 1995 | 1996 | 1997 | death | trees |
| M | M | M | M | M | M | 4 | |
| M | M | M | M | B | M | 1 | |
| M | M | M | M | B | F | 1 | |
| M | M | M | M | N | M | 1 | |
| M | M | M | B | M | M | 1 | |
| M | M | M | B | F | F | 1 | |
| M | M | M | F | B | M | 2 | |
| M | M | M | F | B | N | 1 | |
| M | M | M | F | M | M | 1 | |
| M | M | M | F | F | M | 2 | |
| M | M | M | F | F | F | 1 | |
| M | N | M | M | M | M | 1 | |
| M | N | M | F | M | F | 1 | |
| B | M | M | M | B | M | 1 | |
| B | M | M | F | F | F | 1 | |
| B | B | M | F | F | B | 1 | |
| F | F | B | F | F | F | 1 | |
| F | F | F | F | F | F | 1 | |
| M | M | M | F | F | F | 1997 | 1 |
| B | M | M | F | F | F | 1997 | 1 |
| B | M | M | F | F | B | 1997 | 1 |
| B | B | B | F | F | F | 1997 | 1 |
| F | F | F | F | F | F | 1997 | 1 |
| M | M | M | F | 1995 | 3 | ||
| B | B | B | F | 1995 | 1 | ||
| F | M | M | F | 1995 | 1 | ||
| F | M | F | F | 1995 | 1 | ||
| M | M | M | 1994 | 3 | |||
| M | F | M | 1994 | 1 | |||
| B | M | M | 1994 | 2 | |||
| F | M | M | 1994 | 1 | |||
| M | M | B | 1994 | 2 | |||
| B | F | F | 1994 | 1 | |||
| F | F | F | 1994 | 6 | |||
| B | F | 1993 | 1 | ||||
| F | F | 1993 | 1 | ||||
| F | N | 1993 | 2 | ||||
| B | 1992 | 2 | |||||
| F | 1992 | 2 | |||||
In any given year, except 1995, the majority of males and females maintained the same sex as in the previous year. Most bisexual trees changed to male or female in the next year. In 1995, a drastic sex change towards femaleness occurred synchronously among a large proportion of trees in the population. Out of 35 males in 1994, seven trees (20 %) died and 19 trees (54 %) changed to female or bisexual in 1995 (Table 2). This proportion of sex changes from male to female was significantly higher than that in 1993 and 1994 (Fisher’s exact test with a Bonferroni correction). All sex changes in 1995 of trees that had been bisexual in 1994 were towards femaleness (Table 2). Sex changes from female to male or bisexual were absent in 1995 (Table 2). On the other hand, out of ten females in 1994, seven trees (70 %) died in 1995. However, 20 trees changed to female in 1995 from male or bisexual in 1994. Thus, number and proportion of males in the population decreased to nine trees and 26 %, respectively, in 1995, while number and proportion of females in the population increased to 23 trees and 68 % in 1995.
Table 2.
Frequency of sex changes among Acer rufinerve in each year from 1992–1997
| Sex expression in the next year | |||||||
| Year | n | Male | Bisexual | Female | Non‐flowering | Dead | Relative frequency of sex changes |
| Male | |||||||
| 1992 | 28 | 25 | 0 | 1 | 2 | 0 | 0·04a |
| 1993 | 34 | 31 | 2 | 1 | 0 | 0 | 0·09a |
| 1994 | 35 | 9 | 2 | 17 | 0 | 7 | 0·54b |
| 1995 | 9 | 5 | 3 | 0 | 1 | 0 | 0·33ab |
| 1996 | 8 | 7 | 0 | 1 | 0 | 0 | 0·13ab |
| Bisexual | |||||||
| 1992 | 13 | 6 | 3 | 2 | 0 | 2 | 0·62a |
| 1993 | 3 | 1 | 2 | 0 | 0 | 0 | 0·33a |
| 1994 | 5 | 0 | 0 | 3 | 0 | 2 | 0·60a |
| 1995 | 2 | 1 | 0 | 1 | 0 | 0 | 1·00a |
| 1996 | 6 | 4 | 0 | 1 | 1 | 0 | 0·83a |
| Female | |||||||
| 1992 | 17 | 3 | 0 | 10 | 2 | 2 | 0·18a |
| 1993 | 13 | 1 | 1 | 9 | 0 | 2 | 0·15a |
| 1994 | 10 | 0 | 0 | 3 | 0 | 7 | 0·00a |
| 1995 | 23 | 2 | 3 | 12 | 0 | 6 | 0·22a |
| 1996 | 13 | 2 | 2 | 9 | 0 | 0 | 0·31a |
| Non‐flowering | |||||||
| 1992 | 0 | 0 | 0 | 0 | 0 | 0 | – |
| 1993 | 4 | 2 | 0 | 0 | 0 | 2 | – |
| 1994 | 0 | 0 | 0 | 0 | 0 | 0 | – |
| 1995 | 0 | 0 | 0 | 0 | 0 | 0 | – |
| 1996 | 1 | 1 | 0 | 0 | 0 | 0 | – |
Relative frequencies of plants that changed their sex expression denoted by the same letters do not differ from each other at P < 0·05 (Fisher’s exact test with a Bonferroni correction).
Non‐flowering plants were not analysed statistically because sample sizes were too small.
Out of nine males in 1995, no tree died and three trees (33 %) changed to bisexual and one tree changed to non‐flowering. Three trees changed to male from female or bisexual. Thus, the number of males decreased to eight in 1996. However, the number of females also decreased to 13 in 1996. Out of 23 females in 1995, six trees (26 %) died and five trees changed to male or bisexual. One bisexual changed to female. The male : female ratio increased to 0·62 as a result of a decrease in the number of females, partly by their sex change and partly by their death. Out of eight males in 1996, no tree died and one tree (13 %) changed to female. Seven trees changed to male. Thus, the number of males increased to 14 in 1997. On the other hand, out of 13 females in 1996, no tree died and four trees changed to male or bisexual. Two trees changed to female from male or bisexual. Thus, number of females decreased to 11 in 1997. The male : female ratio increased to 1·27 as a result of sex change.
Inter‐year variation of demographic parameters
The average growth rate in the population was lowest in 1995 (Fig. 4) followed by 1994. The difference in growth rate between years was significant (Kruskal–Wallis test, H = 12·14, n = 5, P < 0·05); however no differences between pairs were detected using a Mann–Whitney test with a Bonferroni correction for multiple comparisons.
Fig. 4. Absolute growth rate (AGR) of stem diameter of Acer rufinerve in each year from 1992 to 1997. The bars represent median of AGR for each year. Median AGRs differed significantly among years (P < 0·05, Kruskal–Wallis test). However, a significant difference at P < 0·05 between pairs was not detected by a Mann–Whitney test with a Bonferroni correction.
A comparison of the AGR of each tree between two successive years showed that the AGR was significantly smaller in 1994 (Wilcoxon test, z = –3·69, n = 50, P < 0·001) and 1995 (z = –4·15, n = 34, P < 0·001) and significantly larger in 1996 (z = –3·77, n = 28, P < 0·001) and 1997 (z = –2·13, n = 28, P < 0·05) than it had been in the previous year.
Over 60 % of A. rufinerve trees in the population died during the study period (Table 1). The annual mortality per year varied from 0 to 0·320 year–1, and differed significantly between years (Table 3). Mortality was significantly higher in 1994 than in 1992, 1993 and 1996 (Fisher’s exact test with a Bonferroni correction).
Table 3.
Comparison of mortality among male, bisexual, female and non‐flowering Acer rufinerve in each year from 1992 to 1997
| Number of plants | ||||
| Sex | Survival | Dead | Total | Mortality |
| 1992 | ||||
| Male | 28 | 0 | 28 | 0·00a |
| Female | 15 | 2 | 17 | 0·12a |
| Bisexual | 11 | 2 | 13 | 0·15a |
| Non‐flowering | 0 | 0 | 0 | – |
| Total | 54 | 4 | 58 | 0·07 |
| 1993 | ||||
| Male | 34 | 0 | 34 | 0·00a |
| Female | 11 | 2 | 13 | 0·15a |
| Bisexual | 3 | 0 | 3 | 0·00a |
| Non‐flowering | 2 | 2 | 4 | 0·50a |
| Total | 50 | 4 | 54 | 0·07 |
| 1994 | ||||
| Male | 28 | 7 | 35 | 0·20a |
| Female | 3 | 7 | 10 | 0·70b |
| Bisexual | 3 | 2 | 5 | 0·40ab |
| Non‐flowering | 0 | 0 | 0 | – |
| Total | 34 | 16 | 50 | 0·32 |
| 1995 | ||||
| Male | 9 | 0 | 9 | 0·00a |
| Female | 17 | 6 | 23 | 0·26a |
| Bisexual | 2 | 0 | 2 | 0·00a |
| Non‐flowering | 0 | 0 | 0 | – |
| Total | 28 | 6 | 34 | 0·18 |
| 1996 | ||||
| Male | 8 | 0 | 8 | 0·00a |
| Female | 13 | 0 | 13 | 0·00a |
| Bisexual | 6 | 0 | 6 | 0·00a |
| Non‐flowering | 1 | 0 | 1 | 0·00a |
| Total | 28 | 0 | 28 | 0·00 |
| 1997 | ||||
| Male | 14 | 0 | 14 | 0·00a |
| Female | 7 | 4 | 11 | 0·36a |
| Bisexual | 1 | 1 | 2 | 0·50a |
| Non‐flowering | 1 | 0 | 1 | 0·00a |
| Total | 23 | 5 | 28 | 0·18 |
Mortalities denoted by the same letters do not differ from each other at P < 0·05 (Fisher’s exact test with a Bonferroni correction).
Differences in demographic parameters between males and females
Most dead trees reproduced as a female before they died. Out of 35 trees that died during the six years, 26 trees (74 %) had female function (21 females and five bisexuals) just before their death and two non‐flowering trees reproduced as female in the year previous to death, while only seven trees (20 %) were males that all died in 1994. This female : male ratio, 21 : 7, was significantly female‐biased (χ2 = 7·00, P < 0·05). The mortality of females or bisexuals occurred every year except 1996, while mortality of males occurred only in 1994 when the population suffered from high mortality. In 1994, females showed higher mortality than males (Fisher’s exact test with a Bonferroni correction; Table 3). In other years, mortality also tended to be higher in females than in males, although these tendencies were not significant.
The AGR of males in 1994 decreased in 1995 when their sex changed to female or to bisexual in 1995 (Fig. 5B; n = 19, P < 0·001), although the AGR of trees that remained male in 1995 did not change (Fig. 5A; n = 9, P > 0·05). The AGR of females in 1995 increased in 1996 when their sex changed to male or to bisexual in 1996 (Fig. 5D; n = 5, P < 0·05), although the growth rate did not change when they maintained their sex (Fig. 5C; n = 12, P > 0·05).
Fig. 5. Comparison of the absolute growth rate (AGR) of stem diameter of Acer rufinerve (A) that reproduced as male both in 1994 and 1995 (n = 9), (B) that reproduced as male in 1994 and as female or bisexual in 1995 (n = 19), (C) that reproduced as female both in 1995 and 1996 (n = 12), and (D) that reproduced as female in 1995 and as male or bisexual in 1996 (n = 5). AGR was compared with a Wilcoxon test.
DISCUSSION
Fluctuation in the sex ratio over 6 years resulted partly from sex changes and partly from mortality. From 1992 to 1994, males outnumbered females, as has been reported for Acer species (Hibbs and Fischer, 1979; Primack and McCall, 1986; Sakai, 1990; Matsui, 1995). However, in 1995, the sex ratio became female‐biased; this has never been reported for Acer species. The male : female ratio became significantly female‐biased in 1995 mainly as a result of sex change to female. The proportion of males that underwent a sex change in 1995 (54 %) was higher than in other years. Such a ‘mass sex change’ has seldom been reported in previous studies on plant species that change their sex (McArthur, 1977; Hibbs and Fischer, 1979; Barker et al., 1982; Primack and McCall, 1986; Kinoshita, 1987; Sakai, 1990; Zimmerman, 1991; Matsui, 1995; Yamashita and Abe, 2002), although Schlessman (1991) reported that 83 % of hermaphroditic Panax trifolium, a perennial herb, became male.
Hibbs and Fischer (1979) have suggested that sex changes from male to female among A. pensylvanicum were induced by a deterioration in growth conditions and plant health, which was caused, in turn, by progressive crown closure in a forest. Matsui (1995) has proposed that a decline in plant health, as a result of water stress, caused male A. rufinerve to become female. These studies provided a hypothetical factor governing sex expression in plants, i.e. plant health, but the process was not indicated. Our results provide evidence to support this hypothesis because a drastic sex change towards increasing femaleness was observed after a decline in growth rate and an increase in mortality within the population. The drastic sex change observed in our study appears to be explained not by a progressive process such as crown closure, but by a sudden process, such as water stress, induced by meteorological conditions. It is quite likely that the growth of A. rufinerve deteriorated and that mortality was higher in 1994 because of drought during the growing season. Drought may be an environmental cue that triggers sex change towards femaleness in A. rufinerve.
The theory of environmental sex determination (Charnov and Bull, 1977) is based on two assumptions: (1) an individual’s fitness as a male or female is influenced by environmental conditions; (2) some environmental conditions favour males, while other conditions favour females. In plants with labile sex expression, males with lower reproductive costs are generally considered to be more advantageous than females with higher reproductive costs under environmental stresses such as low soil moisture, low light intensity, or low levels of nutrients (Freeman et al., 1976, 1980). Thus, we predict that A. rufinerve would change sex towards maleness after experiencing environmental stresses. However, contrary to the prediction, A. rufinerve trees tended to change towards femaleness after the stressful year of 1994.
Dead A. rufinerve trees reproduced as female before they died except dead males in 1994. Changing sex to become female may therefore be a strategy when an individual is near death for A. rufinerve. When a plant dies, a gap is created, and this gap is ultimately occupied by a new plant. Whether a plant that occupies a gap is of the same species as the one that has died is a key factor in maintaining plant populations (Grubb, 1977). Sex change towards femaleness among damaged and dying A. rufinerve trees can be a means by which to disperse seeds before they die. This may facilitate the occupancy by offspring in a gap created by the death of a parent tree. Furthermore, in 1995, all sex‐changed females did not necessarily die. However, it is likely that canopy gaps were created by high mortality of males and females induced by heavy drought in the previous year of 1994. Sex‐changed females could disperse their seeds into the canopy gaps and the sex changes into females might facilitate the regeneration of an Acer rufinerve population, even if the sex‐changed females did not die in 1995. To clarify the role of sex changes in the regeneration of A. rufinerve, seed germination and survival, and the growth of seedlings should be investigated.
ACKNOWLEDGEMENTS
We would like to thank the Kasuga Shinto Shrine for permission to work in their divine forest and Jun Tamai for assistance in establishing the study plot. We are very grateful to Katsuhiko Kimura and Somchai Thoranisorn for their assistance in the field work. We wish to thank Akira Itoh for very helpful comments. This study was partly supported by a grant (12660138, 15580125) from the Ministry of Education, Culture, Sports, Science and Technology of Japan.
Received 30 September 2003; Returned for revision 13 November 2003; Accepted 11 February 2004; Published electronically: 21 April 2004
References
- AbeT.2002. Flower bud abortion influences clonal growth and sexual dimorphism in the understorey dioecious shrub Aucuba japonica (Cornaceae). Annals of Botany 89: 675–681. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Anon. 1994. Annual report of meteorological data in Nara Prefecture. Nara: Nara Meteorological Station, Japan [in Japanese]. [Google Scholar]
- BarkerPA, Freeman DC, Harper KT.1982. Variation in the breeding system of Acer grandidentatum Forest Science 28: 563–572. [Google Scholar]
- BierzychudekP.1982. The demography of jack‐in‐the‐pulpit, a forest perennial that changes sex. Ecological Monographs 52: 335–351. [Google Scholar]
- CharnovEL.1982.The theory of sex allocation. Princeton: Princeton University Press. [PubMed] [Google Scholar]
- CharnovEL, Bull J.1977. When is sex environmentally determined? Nature 266: 828–830. [DOI] [PubMed] [Google Scholar]
- De JongPC.1976.Flowering and sex expression in Acer L A biosystematic study. Wageningen: H. Veenman and Zonen. [Google Scholar]
- FreemanDC, Harper KT, Charnov EL.1980. Sex change in plants: old and new observations and new hypotheses. Oecologia 47: 222–232. [DOI] [PubMed] [Google Scholar]
- FreemanDC, Klikoff LG, Harper KT.1976. Differential resource utilization by the sexes of dioecious plants. Science 193: 597–599. [DOI] [PubMed] [Google Scholar]
- GhiselinMT.1969. The evolution of hermaphroditism among animals. Quarterly Review of Biology 44: 189–208. [DOI] [PubMed] [Google Scholar]
- GrubbPJ.1977. The maintenance of species‐richness in plant communities: the importance of the regeneration niche. Biological Reviews 52: 107–145. [Google Scholar]
- HibbsDE, Fischer BC.1979. Sexual and vegetative reproduction of striped maple (Acer pensylvanicum L.). Bulletin of the Torrey Botanical Club 106: 222–227. [Google Scholar]
- KinoshitaE.1986. Size‐sex relationship and sexual dimorphism in Japanese Arisaema (Araceae). Ecological Research 1: 157–171. [Google Scholar]
- KinoshitaE.1987. Sex change and population dynamics in Arisaema (Araceae). I. Arisaema serratum (Thunb.) Schott. Plant Species Biology 1: 15–28. [Google Scholar]
- KorpelainenH.1994. Sex ratios and resource allocation among sexually reproducing plants of Rubus chamaemorus Annals of Botany 74: 627–632. [Google Scholar]
- LimingFG.1957. Homemade dendrometers. Journal of Forestry 55: 575–577. [Google Scholar]
- LloydDG, Bawa KS.1984. Modification of the gender of seed plants in varying conditions. Evolutionary Biology 17: 255–338. [Google Scholar]
- LloydDG, Webb CJ.1977. Secondary sex characters in plants. Botanical Review 43: 177–216. [Google Scholar]
- Lovett DoustJ, Cavers, PB.1982. Sex and gender dynamics in jack‐in‐the‐pulpit, Arisaema triphyllum (Araceae). Ecology 63: 797–808. [Google Scholar]
- McArthurED.1977. Environmentally induced changes of sex expression in Atriplex canescens Heredity 38: 97–103. [Google Scholar]
- MatsuiK.1991. Pollination ecology of four Acer species in Japan with special reference to bee pollinators. Plant Species Biology 6: 117–120. [Google Scholar]
- MatsuiK.1995. Sex expression, sex change and fruiting habit in an Acer rufinerve population. Ecological Research 10: 65–74. [Google Scholar]
- NanamiS, Kawaguchi H, Yamakura T.1999. Dioecy‐induced spatial patterns of two codominant tree species, Podocarpus nagi and Neolitsea aciculata Journal of Ecology 87: 678–687. [Google Scholar]
- OgataK.1999. Aceraceae. In: Iwatsuki K, Boufford DE, Ohba H, eds. Flora of Japan, Volume IIc Tokyo: Kodansha, 60–73. [Google Scholar]
- OhmaeY, Shibata K, Yamakura T.1996. Seasonal change in nagilactone contents in leaves in Podocarpus nagi forest. Journal of Chemical Ecology 22: 477–489. [DOI] [PubMed] [Google Scholar]
- OrtizPL, Arista M, Talavera S.2002. Sex ratio and reproductive effort in the dioecious Juniperus communis subsp. alpina (Suter) C˘elak. (Cupressaceae) along an altitudinal gradient. Annals of Botany 89: 205–211. [DOI] [PMC free article] [PubMed] [Google Scholar]
- PolicanskyD.1981. Sex choice and the size advantage model in jack‐in‐the‐pulpit (Arisaema triphyllum). Proceedings of the National Academy of Sciences of the USA 78: 1306–1308. [DOI] [PMC free article] [PubMed] [Google Scholar]
- PrimackRB, McCall C.1986. Gender variation in a red maple population (Acer rubrum; Aceraceae): a seven‐year study of a ‘polygamodioecious’ species. American Journal of Botany 73: 1239–1248. [Google Scholar]
- SakaiAK.1978.Ecological and evolutionary aspects of sex expression in silver maple Acer saccharinum. PhD Thesis, University of Michigan, USA. [Google Scholar]
- SakaiAK, Oden NL.1983. Spatial pattern of sex expression in silver maple (Acer saccharinum L.): Morisita’s index and spatial autocorrelation. American Naturalist 122: 489–508. [Google Scholar]
- SakaiAK.1990. Sex ratios of red maple (Acer rubrum) populations in northern lower Michigan. Ecology 71: 571–580. [Google Scholar]
- SakaiS.1987. Patterns of branching and extension growth of vigorous saplings of Japanese Acer species in relation to their regeneration strategies. Canadian Journal of Botany 65: 1578–1585. [Google Scholar]
- SchlessmanMA.1991. Size, gender, and sex change in dwarf ginseng, Panax trifolium (Araliaceae). Oecologia 87: 588–595. [DOI] [PubMed] [Google Scholar]
- SuganumaT, Kawai Y.1978. Phytosociological study of Podocarpus nagi forest in the precincts of Kasuga Shrine, Nara. Journal of Geobotany 25: 221–228 [in Japanese with English summary]. [Google Scholar]
- TatenoR, Kawaguchi H.2002. Differences in nitrogen use efficiency between leaves from canopy and subcanopy trees. Ecological Research 17: 695–704. [Google Scholar]
- UshimaruA, Matsui K.2001. Sex change in tree species: long‐term monitoring of sex expression in Acer rufinerve Nordic Journal of Botany 21: 397–399. [Google Scholar]
- YamashitaA, Abe T.2002. Size distribution, growth and inter‐year variation in sex expression of Bischofia javanica, an invasive tree. Annals of Botany 90: 599–605. [DOI] [PMC free article] [PubMed] [Google Scholar]
- ZimmermanJK.1991. Ecological correlates of labile sex expression in the orchid Catasetum viridiflavum Ecology 72: 597–608. [Google Scholar]